Cool idea, horrible dataset. Its limited to things in Wikidata rather than using OSM, and it must have an article on Wikipedia in English to count so I am sure it is missing just based on that. And somehow it misses many castles (and other categories it includes) which do meet those requirements.
There is no definition of what they use to categorize anything either. And it clearly isn't applied consistently. There is a Château category and yet many things with Château in their name are tagged as castles.
Inferring pronunciation from spelling is very regular in French; the reverse, not so: there may be many ways to spell a sound like “o”, each with a different meaning.
That is an extremely strong claim that I simply cannot support. English is beautiful elegantly designed and trivial to learn compared to french.
Also, if you talk on some mid level (say B1, which takes a LOT of effort for french language), you are not going to get laughed at and ridiculed when using it in UK/US/Australia/NZ etc with english. A very common experience in France when trying to use french (I go there every other weekend, its just across the border). Folks from other parts of former french empire like quebecois or corsicans are much more approachable in this and tolerant. Suffice to say their language variant is often mocked by mainland french too, I think personally parisians are the most cocky ones.
But you don’t pronounce it and in your example you know it’s plural without the x.
I would say it’s technical debt but rest assured that the language is alive and will eventually drop the x. Perhaps we will have a s instead of a x first.
There is no definition of what they use to categorize anything either. And it clearly isn't applied consistently. There is a Château category and yet many things with Château in their name are tagged as castles.